[aarch64-port-dev ] RFR: 8172144: AArch64: Implement "JEP 270: Reserved Stack Areas for Critical Sections"

Bob Vandette bob.vandette at oracle.com
Sat Mar 18 14:24:26 UTC 2017

The way we currently exclude tests in jtreg is to use @requires and provide specific os names or architectures.
I really don’t want to have 2 different os.arch names for the two 64-bit arm ports.  This would cause issues for 
developers switching off of the os.arch names.   So unless we extend jtreg (which we could do at some point)
to provide additional variables to test for,  then we need to stick with the variables that exist.  

It perfectly acceptable for a test to be written in such a way as to verify that a feature exists and return success
when the feature is not available.  That doesn’t appear to be the case in the webrev that you posted.

ps.  This is just my .02, I’m not the grand Poobah for these types of issues.


> On Mar 18, 2017, at 4:15 AM, Andrew Haley <aph at redhat.com> wrote:
> On 17/03/17 21:01, Bob Vandette wrote:
>> I’m just saying that the test should run on both the Oracle 64-bit
>> ARM port as well as the aarch64 port and shouldn’t be excluded on
>> one or the other implementations.
> But you haven't explained why you believe this.  It's not practical to
> keep both implementations in lock-step with regard to features.  At
> least, not without hobbling one of them, which I would find totally
> unacceptable.  And we don't want tests enabled on platforms which
> don't support a feature under test, do we?
> Andrew.

More information about the hotspot-runtime-dev mailing list